Binding of Hepatitis C Virus to CD81

See allHide authors and affiliations

Science  30 Oct 1998:
Vol. 282, Issue 5390, pp. 938-941
DOI: 10.1126/science.282.5390.938


Chronic hepatitis C virus (HCV) infection occurs in about 3 percent of the world's population and is a major cause of liver disease. HCV infection is also associated with cryoglobulinemia, a B lymphocyte proliferative disorder. Virus tropism is controversial, and the mechanisms of cell entry remain unknown. The HCV envelope protein E2 binds human CD81, a tetraspanin expressed on various cell types including hepatocytes and B lymphocytes. Binding of E2 was mapped to the major extracellular loop of CD81. Recombinant molecules containing this loop bound HCV and antibodies that neutralize HCV infection in vivo inhibited virus binding to CD81 in vitro.

HCV is a positive strand RNA virus of the flaviviridae family (1) chronically infecting about 170 million persons worldwide (2). Chronic HCV infection results in liver diseases (hepatitis, cirrhosis, and hepatocellular carcinoma) in a sizable fraction of cases (3). Infection with HCV is also associated with most cases of type II and type III cryoglobulinemia, B lymphocyte proliferative disorders characterized by polyclonal B cell activation and autoantibody production (4). The complete HCV sequence has been available since 1989 (5); however, progress in understanding the viral life cycle has been hampered by the lack of virus culture systems in vitro. Although hepatocytes and B lymphocytes are thought to be infected by HCV (1), there is no consensus on viral tropism, and the cellular receptor for the virus has not been identified.

We have shown previously that a recombinant form of the major envelope protein (E2) of HCV binds with high affinity to human lymphoma and hepatocarcinoma cell lines, whereas it does not bind to mouse cells (6). Furthermore, in chimpanzees vaccinated with recombinant E1 and E2 envelope proteins, protection from homologous HCV challenge correlated with the presence of antibodies capable of inhibiting the binding of E2 to human cells (6). These results suggested that E2 may be responsible for binding of HCV to target cells.

To identify the E2-binding molecule on human cells, we prepared a cDNA expression library from A2R cells, a subclone of the human T cell lymphoma Molt-4, which exhibited high E2-binding capacity (7). This library was screened by transient transfection of a mouse fibroblast cell line (WOP) (8) with E2 as a probe (9). This approach resulted in the identification of a cDNA clone that conferred E2-binding capacity to WOP cells upon transient transfection. This clone contains an insert of 0.9 kb encoding human CD81. This widely expressed 25-kD molecule is a member of the tetraspanin superfamily (10), which includes cell surface proteins that span the membrane four times, forming two extracellular loops. The intracellular and transmembrane domains of CD81 are highly conserved among different species, whereas the major extracellular loop is quite diverse (Fig. 1). The major loop is highly conserved in humans and chimpanzees, which are the only known species permissive to HCV infection (1) and whose cells bind HCV E2 (6).

Figure 1

Expression cloning of the HCV E2-binding molecule. Alignment of CD81 amino acid sequences from human (A2R cell line), chimpanzee (peripheral blood mononuclear cells) (26), green monkey (COS cell line) (26), hamster (CHO cell line) (26), rat (27), and mouse (28). Predicted cytoplasmic (CY), transmembrane (TM), and extracellular (EC) domains are indicated according to (10). Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.

To confirm that CD81 was the human cell surface molecule binding HCV E2, we used recombinant E2 and antibodies to CD81 (anti-CD81) to assess this interaction. Recombinant E2 competitively inhibited the binding of anti-CD81 to Epstein-Barr virus (EBV)–transformed B cell lines (EBV-B cells) (Fig. 2A). In addition, E2 reacted in protein immunoblots with anti-CD81–precipitated material (Fig. 2B). CD81 on fresh lymphocytes and hepatocytes is also capable of binding E2, as demonstrated by immunohistochemical staining with biotin-labeled E2 (11).

Figure 2

Interaction between recombinant E2 and CD81. (A) Dose-dependent inhibition of anti-CD81 binding to B cells by recombinant E2. EBV-B cells were incubated with increasing concentrations of recombinant E2 for 1 hour at 4°C and then stained with an mAb to CD81 (clone JS-81, Pharmingen). The data are expressed as percentage of inhibition of mean fluorescence intensity (6). Preincubation with E2 did not inhibit binding of an anti–major histocompatibility complex class I (clone W6/32) (29). (B) E2 detects, on protein immunoblot, the 25-kD protein immunoprecipitated by anti-CD81. A membrane protein extract (about 300 μg) prepared from the A2R cell line was solubilized with 8 mM 3-[(3-cholamidopropyl)-dimethyl-ammonio]-1-propanesulfonate in PBS (pH 7.4) and incubated with 10 μg of recombinant E2 (lanes 2 and 3), 20 μg of anti-CD81 (lane 4), or 20 μg of a control antibody (anti–human CD9; ATCC) (lane 5). After incubation for 2 hours at 4°C, the samples were immunoprecipitated with chimpanzee antisera to E2 (lane 2), chimpanzee preimmune sera (lane 3), or goat anti–mouse IgG (lanes 4 and 5) bound to protein A–Sepharose (CL-4B; Pharmacia). The precipitated samples were eluted in Laemmli buffer, separated by SDS-PAGE under nonreducing conditions, and transferred to a PVDF (polyvinylidene difluoride) membrane by electroblotting. The blots were then probed with recombinant E2 (1 μg/ml) overnight followed by a 2-hour incubation with an mAb to E2 (mAb 291). Visualization of the immunoprecipitated proteins detected by E2 was performed with a peroxidase-conjugated polyclonal anti–mouse IgG (Amersham). As a positive control, a portion of the total membrane extract was also loaded on the gel (lane 1). The mobilities of molecular mass (in kilodaltons) markers are indicated at left. The 25-kD CD81 is immunoprecipitated directly by anti-CD81 or indirectly by a combination of E2 and anti-E2 as shown in lanes 4 and 2, respectively. The high molecular mass bands observed in all lanes (including negative control lane) probably represent cross-reaction of the secondary antibody, detection of E2 precipitated by chimpanzee antisera to E2 subsequently recognized by anti-E2 mAb (lane 2), or detection of mAb to CD81 (lane 4) or mAb to CD9 (lane 5) by the secondary anti-mouse.

Because of the lack of HCV culture assays in vitro, we developed alternative methods to demonstrate that CD81 interacts with HCV. We prepared a recombinant fusion protein (Fig. 3A) expressing the major extracellular loop (EC2) (amino acid residues 113 to 201) of human or mouse CD81 fused to the COOH-terminal end of thioredoxin (TRX-EC2) (12). The proteins containing the human, but not the mouse, loop bound to E2 in protein immunoblot (13) and in solution as shown by inhibition of binding of E2 to human cells (Fig. 3B). To assess virus binding, we attached human or mouse TRX-EC2 proteins to polystyrene beads and incubated them with an infectious plasma containing known amounts of viral RNA molecules. After washing, the amount of bead-associated virus was measured by quantitative reverse transcription polymerase chain reaction (RT-PCR) (14). Beads coated with human TRX-EC2, but not with mouse TRX-EC2, bound HCV in a concentration-dependent fashion (15) (Fig. 3C). Preincubation of beads with anti-CD81 inhibited virus binding. Furthermore, preincubation of the infectious plasma with sera from chimpanzees that were protected from homologous HCV challenge by vaccination with E1 and E2 envelope proteins (16) inhibited HCV binding to CD81 (Fig. 3D). In contrast, sera from vaccinated but nonprotected chimpanzees, although containing anti-E2, were not inhibitory (Fig. 3D). These results demonstrate that anti-E2 antibodies, which are capable of neutralizing HCV infection in vivo, can inhibit the binding of HCV to CD81 in vitro, supporting the idea that CD81-E2 interaction is relevant to infection.

Figure 3

The major extracellular loop of CD81 binds recombinant E2 and viral particles. (A) Schematic representation of the pThioHis-EC2 plasmid in the region encoding the thioredoxin-EC2 fusion protein. The nucleotide and the amino acid sequences of the joining region are indicated. The enterokinase cleavage site is underlined. Plac, lac promoter; SD, ribosome-binding site; TRX, thioredoxin; EC2, human major extracellular loop (amino acid residues 113 to 201) of CD81. (B) Dose-dependent inhibition of E2 binding to hepatocarcinoma cells by recombinant molecules expressing the EC2 of human CD81. Cells were incubated with E2 and increasing concentrations of human (closed circles) or mouse (open circles) TRX-EC2 for 1 hour at 4°C and then stained with mAb to E2. Data are expressed as percentage of inhibition of mean fluorescence intensity. (C) Binding of HCV to CD81. Polystyrene beads were coated with human (hEC2) or mouse (mEC2) TRX-EC2 (10 or 50 μg/ml) overnight at room temperature (14). Each bead was then incubated at 37°C with chimpanzee infectious plasma (genotype 1a) containing 5 × 105HCV RNA molecules in 200 μl. The bound virus was eluted with lysis buffer, and HCV RNA was measured by quantitative RT-PCR (14). Similar results were obtained with an infectious plasma containing HCV of the genotype 1b. For inhibition experiments, the TRX-EC 2–coated beads were incubated with mAb to CD81 or with an isotype-matched irrelevant antibody [hepatitis B surface antigen (anti-HBsAg)] as control, for 1 hour at room temperature before incubation with the virus (14). (D) Antibodies that neutralize HCV infection in vivo inhibit binding of HCV to CD81 in vitro. The chimpanzee infectious plasma (200 μl) used in the experiment described in (C) was preincubated at 4°C for 1 hour with 2 μl of serum either from a chimpanzee (number 559) protected from homologous HCV challenge by vaccination with E1 and E2 envelope proteins or from a chimpanzee (number 590) not protected from homologous HCV challenge after vaccination with E1 and E2 envelope proteins (6). As a control, chimpanzee infectious plasma was preincubated with preimmunization sera. The plasma was then incubated with beads that were coated with human TRX-EC2 (10 μg/ml), and HCV binding was assessed as described in (C). Results similar to that of serum from chimp 559 were obtained with sera from three other chimpanzees (numbers 357, 534, and 653) that were vaccinated and protected (6, 16).

Our data demonstrate that human CD81 is sufficient for binding not only E2 but also HCV particles. Given the wide distribution of CD81 (10), these results imply that HCV can bind to a variety of cells other than hepatocytes. Consistent with this finding, HCV RNA has been found in T and B lymphocytes and monocytes (17). Whether virus binding to CD81 is followed by entry and infection in all cell types is not clear, because it is possible that additional factors are required for HCV fusion or infectivity.

CD81 participates in different molecular complexes on different cell types, a fact that may affect its capacity to mediate HCV attachment or to deliver signals to target cells. For instance, on epithelial and hematopoietic cells, CD81 associates with integrins (18), whereas on B cells it associates with CD21 and CD19 (19), forming a complex that, when appropriately engaged, can lower B cell activation threshold (20). EBV and HCV target this complex by binding CD21 (21) and CD81, respectively. The B cell activation capacity of EBV is well known, and we have evidence that the HCV envelope protein E2 delivers a costimulatory signal to B cells (22). It may well be that the binding of HCV to CD81 on B lymphocytes in vivo lowers the activation threshold of these cells, thus facilitating the production of autoantibodies that are the hallmark of HCV-associated cryoglobulinemia (4).

Identification of the interaction between CD81 and HCV could help to elucidate the pathogenesis of HCV-associated diseases, obtain a small animal model of infection, and develop new therapeutic strategies directed at interfering with virus binding.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: abrignani{at}


View Abstract

Stay Connected to Science

Navigate This Article